摘要
本文首先介绍了网络仿真技术的概念、特点和发展现状,在此基础上介绍了当前主流的通用型网络仿真软件的主要特点并对它们进行了对比分析和评价.为了解决现有网络架构灵活性和服务能力差的问题,我们引入了面向大数据建模、能够对网络整体状态进行预测和实时优化的平行网络架构及其计算实验平台.该平台融合虚拟的人工网络和实际网络的数据并对其进行实验分析和评估,可以发现最优的控制方案.基于平行网络计算实验方法,本文设计了针对微信朋友圈转发预测的网络优化方法,并通过实验验证了该方法可以有效提升网络的性能.
In this paper, the current network experiment platforms are investigated, introduced and compared. The most popularly used simulator and its main advantages and disadvantages are demonstrated. To solve the problem of lacking flexibility and service capability of traditional network experiment platforms, based on the parallel network architecture, the corresponding computational experiment platform is proposed. The proposed platform is data-oriented, and by using the computational experiments and analysis, an optimized control strategy can be continuously updated and tracked, thus the self optimization of network systems is achieved. In the end, a computational experiment based on retweet analysis of Wechat Moments is proposed, and the effectiveness of this method is evaluated.
作者
杨林瑶
韩双双
王晓
李玉珂
王飞跃
YANG Lin-Yao;HAN Shuang-Shuang;WANG Xiao;LI Yu-Ke;WANG Fei-Yue(The State Key Laboratory for Management and Control of Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing 100190;Qingdao Academy of Intelligent Industries, Qingdao 266000;University of Chinese Academy of Sciences, Beijing 100049;Vehicle Intelligence Pioneers Inc., Qingdao 266000;Research Center of Military Computational Experiments and Parallel Systems Technology, National University of Defense Technology, Changsha 410073)
出处
《自动化学报》
EI
CSCD
北大核心
2019年第9期1637-1654,共18页
Acta Automatica Sinica
基金
国家自然科学基金(61501461,61702519,61533019,91720000)
北京市科委项目(Z181100008918007)
英特尔智能网联汽车大学合作研究中心(ICRI-IACV)资助~~
关键词
仿真软件
网络仿真
平行网络
计算实验
虚实交互
Simulators
network simulation
parallel network
computational experiments
virtual-reality interaction
